Adjustment of Intention to Start a Motion of Lower
Limbs Based on Cerebral Hemoglobin Information
Chunguang Li, Juan Li, Haiyan Hu, Weida Li
Key Laboratory of Robotics and System of Jiangsu Province
School of Mechanical and Electric Engineering, Soochow
University
Suzhou, China
E-mail: lichunguang@suda.edu.cn
, lijuan_722@163.com
Wei Qu
Project Management Department
ECOVACS Robotics
Suzhou, China
Abstract—To increase intelligence of walking-assisted devices,
it is important to provide a start command of action based on the
motion intention of subjects. In this paper, a method for
identifying motion intention was proposed based on cerebral
hemoglobin information. Spontaneous movements (upstairs,
downstairs, upslope, down-slope, sit-down, squat, and the
corresponding standup) without an advanced imagination or
external stimuli were performed on two subjects. During the
experiment, cerebral hemoglobin information was recorded by
applying near-infrared spectroscopic technology. Multiple
analyses of variance on different features of the variation of total
hemoglobin were combined and a series of significant levels were
set. The corresponding precision rate was up to 74.8%, and the
sensitivity was up to 79.2%. Since the proposed identification
method is realized based on the brain information recorded in a
real movement environment, it is helpful and practical to
enhance intelligence of walking-assisted devices.
Keywords—motion intention; identification; statistical analysis;
variation rate and acceleration; spontanous movement
I. INTRODUCTION
Recently, the number of persons with motor dysfunction in
lower limbs has been increasing quickly because of natural
disasters, injuries, accidents, diseases, and so on. In order to
restore patients' ability of independent walking, there is an
urgent demand for intelligent walking-assisted devices. This
requires walking-assisted devices to provide auxiliary power
when patients have a motion intention.
Traditionally, motion intention of a patient is recognized
based on biomechanical information [1–3]. However, some
involuntary trembles of patients affect the adjustment of a real
motion intention. Recently, to control the movement of a
device by using brain information has been attracting much
attention in worldwide. Israel and Bar-Ilan University [4] used
unctional magnetic resonance imaging (fMRI) technology to
identify a subject's intentions and controlled a humanoid robot
to convert the intentions into actions. The subject imagined the
movements of left hand, the right hand, or the legs, and the
robot performed the motions of left, right, or walk forward,
respectively. Gutiérrez-Flores [5] detected motion intent of a
person and controlled a mobile device to execute a command.
The detection rate was up to 95% with 0.2 associated noise.
Honda automobile research and development subsidiary etc.
[6] recognized brain intention by analyzing brain waves and
blood flow changes in brain together. And they controlled the
robot "Asimo" to raise a hand or a leg based on the brain
intention, the corresponding recognition rate is around 90%.
Nagaoka University of Technology [9] estimated three levels
of force strength of a arm based on the changes in blood
oxygen concentration. Meanwhile, a switch command
corresponding to the start and end of a movement could be
given too. Identification rate reached more than 72% on
average for four subjects. However, in the most researches [7–
8] that recognized motion intention by applying brain
information, the subjects should remain in a static state
without any movement. And the identification of motion
intention was based on the imagination information of the
brain. Moreover, in some researches [5–8] an external stimuli
was required for pattern recognition. However, similar
researches to identify motion intention with an actual
movement of lower limbs are very few.
In order to provide a switch command for a walking-
assisted device intelligently, a method for recognizing motion
intention by using brain information of the subjects who
perform a real action of lower limbs was proposed. Compared
to the technologies of magneto encephalography (MEG),
electroencephalography (EEG), and fMRI, near infrared
spectroscopy (NIRS) has fewer restrictions on the testing
environment and subjects. Therefore, NIRS technology was
applied for recording brain information. And motion intention
was identified based on the recorded cerebral hemoglobin
information.
II. E
XEPERIMENTS
A. Subjects
Two healthy volunteers (males, 25 years old) at Soochow
University participated in the motion experiment. No
volunteer had a history of physical or psychiatric disorders.
Informed consent was obtained from each volunteer.